EP0922251B1 - Systeme de memoire - Google Patents
Systeme de memoire Download PDFInfo
- Publication number
- EP0922251B1 EP0922251B1 EP97936594A EP97936594A EP0922251B1 EP 0922251 B1 EP0922251 B1 EP 0922251B1 EP 97936594 A EP97936594 A EP 97936594A EP 97936594 A EP97936594 A EP 97936594A EP 0922251 B1 EP0922251 B1 EP 0922251B1
- Authority
- EP
- European Patent Office
- Prior art keywords
- memory
- bits
- word
- error
- bit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Lifetime
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/08—Error detection or correction by redundancy in data representation, e.g. by using checking codes
- G06F11/10—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's
- G06F11/1008—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's in individual solid state devices
- G06F11/1012—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's in individual solid state devices using codes or arrangements adapted for a specific type of error
- G06F11/1016—Error in accessing a memory location, i.e. addressing error
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/08—Error detection or correction by redundancy in data representation, e.g. by using checking codes
- G06F11/10—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's
- G06F11/1008—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's in individual solid state devices
- G06F11/1012—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's in individual solid state devices using codes or arrangements adapted for a specific type of error
- G06F11/1028—Adjacent errors, e.g. error in n-bit (n>1) wide storage units, i.e. package error
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
- G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
- G11B20/10—Digital recording or reproducing
- G11B20/18—Error detection or correction; Testing, e.g. of drop-outs
- G11B20/1833—Error detection or correction; Testing, e.g. of drop-outs by adding special lists or symbols to the coded information
Definitions
- EDC error detection code
- EDC controllers The formation of the K bits (EDC coding), the comparison, the Decoding of the syndromes, as well as the correction and possibly A higher-level controller is alerted today usually with the help of special controller modules that hereinafter also referred to as EDC controllers.
- 1 shows an EDC code table, About which bit positions of a data word the K bits in an EDC controller.
- the sign "0" next to the lines of the lower half of the memory mean that the associated C bit is equal to 1 if the number of the in the Parity formation included "len” in the entire useful bit part is odd.
- the letter “E” next to the lines of the lower one Memory value half means that the associated C bit is the same Is 0 if the number of those involved in parity formation "len” is odd in the entire useful bit part. The two the latter statements therefore relate to both halves of the memory word together.
- data words are required, which comprise 32 data bits.
- Eight control bits C0, C1, C2 .... C7 assigned, each through parity formation via certain bit positions of a data word are formed.
- the entire memory word i.e. the useful word (address or data) plus control word, therefore comprises 40 bits. These are in DRAM memory devices with a cell width of four bits organized.
- one-bit errors can be reliably detected and lead to odd-numbered syndrome patterns. Furthermore, one Correction of one-bit errors can be done as everyone Error bit position within a useful word is unique Syndrome pattern is firmly assigned. This syndrome pattern can decoded and thus used to correct the faulty bit become.
- Double bit error always lead to an even number, for example Syndrome patterns are not equal to 0 and are therefore considered to be multi-bit errors recognized.
- the other even multi-bit errors also always lead to even-numbered syndrome patterns, whereby with a probability of 1/128 the zero syndrome arises because with this ECC width (number of K bits) there are a total of 128 even-numbered syndrome patterns. These mistakes therefore lead immediately with a probability of 99.2% Multi-bit error alarm.
- This also takes into account the even-numbered multi-bit errors, which cause zero syndrome in 1 of 128 cases.
- FIG. 2 shows what is known as an EDC code table, Which bit positions of an address word the K bits in an EDC controller. For the explanation of the Representation in FIG. 2 is the same as in FIG. 1.
- the problem mentioned can be significantly reduced if the storage units (e.g. memory modules) that are used for reading can be activated together by several control signals from the same type, that of independent control units come, is supplied. In this case, you only count on failure of one of these signals, e.g. Dates and Control bits of different memory words when reading out can be mixed together.
- this measure which is abbreviated as measure A) below, can still, however, with less Probability, one-bit error or even error-free to be faked.
- measure B The clever division of the data and control bits between the Storage units in connection with the associated choice of EDC codes are abbreviated below as measure B).
- the invention is based on the object of being recognizable to improve the latter multi-bit errors.
- An embodiment of the invention is defined in claim 2 specified. This embodiment has the advantage that the detectability of multi-bit errors is still essential is improved.
- An embodiment of the invention is defined in claim 3 specified. This embodiment has in its combination with claim 2 the effect that the detectability of multi-bit errors with almost 100% security is guaranteed.
- FIG. 1 shows an EDC code which is used for Implementation of measure B) for a storage system with two Storage units is selected accordingly.
- the EDC code shown in Fig. 1 is such chosen that a data word section, i.e. a section of the Data word that is not related to another section of the data word is saved (e.g. sections DWT1 or DWT2 in Fig. 1), on the formation of a K bit with each an even number of bit positions is involved, insofar as it is a section of data word that is not together is saved with this K bit, and with an odd one Number of bit positions involved, insofar as it is one Data word section acts together with this K bit is saved.
- a data word section i.e. a section of the Data word that is not related to another section of the data word is saved (e.g. sections DWT1 or DWT2 in Fig. 1), on the formation of a K bit with each an even number of bit positions is involved, insofar as it is a section of data word that is not together is saved with this K bit, and with an odd one Number of bit positions involved, insofar as it is
- FIG. 3 shows a memory system CMY according to the invention, the one double (storage) control system CMYC, one (storage) Medium system CMYM (short "storage”) with two Storage units, i.e. two (storage) medium halves, as well a double BCMY bus interface, via which the Connection to the bus system of a computer system established becomes.
- the storage system includes an error monitor that with the help of EDC controllers depending on the Machine cycle (read or write cycle) provides different checkpoints.
- FIG 3 shows in particular the configuration of the data paths and Control points of the storage system for one write cycle shown.
- the write data and associated data arrive during the write cycle K bits from the BCMY are initially received via write paths WDPI a checkpoint A1.
- the already mentioned error monitoring carried out at the Bit errors are recognized and One-bit errors can even be corrected.
- At checkpoint B i.e. immediately before enrolling the Write data to memory is in each of the two Transfer halves from the write data again Control word generated and an ECC comparison of the two Control words carried out. Only if the data is in Both halves have been processed without errors, are the two Control words are identical, otherwise there is an alarm a multiple error, i.e. one that cannot be corrected Error.
- the ECC comparison finds errors that i.a. on a faulty transmission between A1 and B based or through faulty EDC networks to the Checkpoints A1 or B are conditional.
- control bits of control point A1 are for the Checkpoint B, not needed because it is at checkpoint B, i.e. immediately before writing the write data into Medium system from which write data is generated again and because they are not used for error monitoring between A1 and B. are needed.
- the error monitoring between A1 and B is yes, already by comparing the two at checkpoint B formed control word already guaranteed.
- the data including the control word are transferred to the medium system registered.
- the data bits 0 ... 15 (DWT1) together with K-Bits 4 ... 7 (KWT2) in the a medium half is written and the data bits 16 ... 32 (DWT2) together with the K bits 0 ... 3 (KWT1) in the other Medium half.
- the in section 4 i.e. is done directly upon registration instead of a pure write cycle, a read / write cycle carried out, an EDC check of the read data and thus a Check the control of a memory cell before a correctly controllable memory cell by the Incorrect control of a faulty write process is overwritten.
- the reading process remains internal, i.e. the BCMY does not record the read data in order to transfer them to the To issue bus.
- the write data will be kept until the end of the Read checks cached in the CMYM. Based on these Measure the system effects of control errors kept low, and the (control) errors better distinguished and recognized faster.
- EDC coding has so far been used explained only with regard to the data word. Indeed not only the data word but also the Memory address of the data word in the EDC coding, i.e. in the parity formation included (see FIG. 2).
- the EDC code is chosen with respect to the address word that with a one-bit error in the address word only one even-numbered syndrome pattern can arise. That means but that a multi-bit error is also just one even-numbered syndrome pattern can arise. So it comes in the event of an addressing error, always alarms a Multiple errors, i.e. an uncorrectable error.
- the Including the address in the EDC coding is in the Publication DE 3319710 A1 (internal to SAG 83 P 1382 DE) explained in more detail and is abbreviated below as a measure C) designated.
- the data word and the associated control word is first read out and reaches the Checkpoint C, at which an error detection and if necessary, error correction is carried out. Then the data word and the control word become the BCMY bus interface forwarded at the checkpoint A2 error detection is carried out again.
- error detection causes the Bus interface BCMY the transfer of the data word and Control word on the bus blocked.
- the defective storage system half (CMY0 or CMY1) of the BCMY bus interface switched off.
- FIG. 5 shows in conjunction with FIGS. 3 and 4 at which control point (Checkpoint Y) data errors can be detected in a specific section (Section X) of the storage system occur.
- Control 0 incorrectly read from address 1 when reading at address 2. This leads to the following reading information:
- Control 1 incorrectly switched the associated data and ECC outputs to low resistance when reading at address 1. This leads to the following reading information:
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
- Detection And Correction Of Errors (AREA)
Claims (3)
- Système de mémoire comportanta) une mémoire (CMYM) dans laquelle sont mis en mémoire des mots de mémoire qui comprennent chacun un mot de données et un mot de contrôle,a1) la mémoire comprenant deux unités de mémorisation et chaque fois une partie du mot de données étant mise en mémoire dans chaque unité de mémorisation conjointement avec une partie du mot de contrôle,a2) et la mémoire comprenant deux dispositifs de surveillance d'erreurs qui effectuent à l'aide du mot de contrôle une surveillance d'erreurs du mot de mémoire,b) une commande (CMYC) de mémoire pour commander la mémoire, qui comprend également deux dispositifs de surveillance d'erreurs pour la surveillance d'erreurs entre la commande (CMYC) de mémoire et la mémoire (CMYM),c) une structure de ligne doublée entre la commande (CMYC) de mémoire et la mémoire (CMYM) pour le transfert doublé des mots de mémoire entre la commande (CMYC) de mémoire et la mémoire (CMYM),
- Système de mémoire suivant la revendication 1, caractérisé en ce que les dispositifs de surveillance d'erreurs mentionnés effectuent la surveillance d'erreurs de telle manière qu'ilsa) produisent au moyen d'un dispositif de codage, suivant une loi de formation déterminée, un mot de contrôle à partir du mot de mémoire à surveiller, comparent les bits de ce mot de contrôle (bits K) aux bits K contenus dans le mot de mémoire et, en cas d'inégalité, déduisent le genre de l'erreur du modèle des bits K égaux et inégaux, de ce que l'on appelle le modèle de syndrome,b) la loi de formation mentionnée étant choisie de telle manière que la comparaison précitée donne, pour une erreur sur un bit, un nombre impair de bits K inégaux, chaque fois un nombre pair de bits K inégaux contribuant au nombre impair obtenu à partir des parties du mot de contrôle qui ne sont pas mémorisées conjointement avec la partie du mot de données dans laquelle se trouve l'erreur sur un bit.
- Système de mémoire suivant la revendication 1 ou 2, caractérisé en ce que la subdivision par parties mentionnée des bits de données et des bits K sur les deux unités de mémorisation du système de mémoire est (pour un code EDC prescrit) choisie de telle manière que participent à un modèle de syndrome impair reposant sur une erreur sur un bit un nombre pair de bits K qui ne sont pas mémorisés conjointement avec la partie du mot de données dans laquelle se trouve l'erreur sur un bit.
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE19635237A DE19635237A1 (de) | 1996-08-30 | 1996-08-30 | Speichersystem |
DE19635237 | 1996-08-30 | ||
PCT/DE1997/001644 WO1998009216A1 (fr) | 1996-08-30 | 1997-08-06 | Systeme de memoire |
Publications (2)
Publication Number | Publication Date |
---|---|
EP0922251A1 EP0922251A1 (fr) | 1999-06-16 |
EP0922251B1 true EP0922251B1 (fr) | 2000-05-24 |
Family
ID=7804212
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP97936594A Expired - Lifetime EP0922251B1 (fr) | 1996-08-30 | 1997-08-06 | Systeme de memoire |
Country Status (6)
Country | Link |
---|---|
US (1) | US6539504B1 (fr) |
EP (1) | EP0922251B1 (fr) |
JP (1) | JP2000517080A (fr) |
AR (1) | AR010217A1 (fr) |
DE (2) | DE19635237A1 (fr) |
WO (1) | WO1998009216A1 (fr) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6981091B2 (en) * | 2001-10-25 | 2005-12-27 | Xilinx,Inc. | Using transfer bits during data transfer from non-volatile to volatile memories |
US20050277873A1 (en) * | 2004-05-27 | 2005-12-15 | Janice Stewart | Identification information recognition system for a medical device |
US8961461B2 (en) * | 2004-05-27 | 2015-02-24 | Baxter International Inc. | Multi-state alarm system for a medical pump |
US7927313B2 (en) * | 2004-05-27 | 2011-04-19 | Baxter International Inc. | Medical device configuration based on recognition of identification information |
US7203890B1 (en) | 2004-06-16 | 2007-04-10 | Azul Systems, Inc. | Address error detection by merging a polynomial-based CRC code of address bits with two nibbles of data or data ECC bits |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE3528902C2 (de) * | 1984-12-13 | 1994-06-09 | Siemens Ag | Durch einen EDC-Code gesichertes Speichersystem |
US5043990A (en) * | 1987-12-04 | 1991-08-27 | Hitachi, Ltd. | Semiconductor integrated circuit device |
DE58909568D1 (de) * | 1988-08-02 | 1996-02-22 | Siemens Ag | Verfahren zur Fehlersicherung in Speichersystemen von Datenverarbeitungsanlagen, insbesondere Fernsprechvermittlungsanlagen |
US4993028A (en) * | 1988-09-07 | 1991-02-12 | Thinking Machines Corporation | Error detection and correction coding |
US5872780A (en) * | 1992-05-21 | 1999-02-16 | Alcatel Network Systems, Inc. | Sonet data transfer protocol between facility interfaces and cross-connect |
US5379415A (en) * | 1992-09-29 | 1995-01-03 | Zitel Corporation | Fault tolerant memory system |
KR960000681B1 (ko) * | 1992-11-23 | 1996-01-11 | 삼성전자주식회사 | 반도체메모리장치 및 그 메모리쎌 어레이 배열방법 |
US5907717A (en) * | 1996-02-23 | 1999-05-25 | Lsi Logic Corporation | Cross-connected memory system for allocating pool buffers in each frame buffer and providing addresses thereof |
-
1996
- 1996-08-30 DE DE19635237A patent/DE19635237A1/de not_active Withdrawn
-
1997
- 1997-08-06 JP JP10511150A patent/JP2000517080A/ja active Pending
- 1997-08-06 WO PCT/DE1997/001644 patent/WO1998009216A1/fr active IP Right Grant
- 1997-08-06 US US09/230,757 patent/US6539504B1/en not_active Expired - Fee Related
- 1997-08-06 DE DE59701775T patent/DE59701775D1/de not_active Expired - Fee Related
- 1997-08-06 EP EP97936594A patent/EP0922251B1/fr not_active Expired - Lifetime
- 1997-08-29 AR ARP970103951A patent/AR010217A1/es active IP Right Grant
Also Published As
Publication number | Publication date |
---|---|
WO1998009216A1 (fr) | 1998-03-05 |
AR010217A1 (es) | 2000-06-07 |
DE59701775D1 (de) | 2000-06-29 |
EP0922251A1 (fr) | 1999-06-16 |
DE19635237A1 (de) | 1998-03-05 |
US6539504B1 (en) | 2003-03-25 |
JP2000517080A (ja) | 2000-12-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP0922254B1 (fr) | Reconnaissance des erreurs dans un systeme de memoire | |
DE60001370T2 (de) | Verfahren und vorrichtung zur erkennung von doppelbitfehlern und korrektur von fehlern durch bauelementfehler verursacht | |
DE3111447C2 (fr) | ||
DE3882223T2 (de) | Ausgebreitete Fehlerkorrekturvorrichtung mit Einzel-Paket-Fehlerkorrektur und Doppel-Paket-Fehlerdetektionscoden. | |
DE2260850A1 (de) | Fehlerkorrektursystem | |
DE2357168A1 (de) | Speichermodul fuer eine datenverarbeitungseinheit | |
DE69904618T2 (de) | Detektionstechnik von speicherabschnittfehlern und einzel-, doppel und triplebitfehlern | |
EP0141160A2 (fr) | Montage pour l'enregistrement d'adresses de cellules de mémoire à contenus erronés | |
EP0922253B1 (fr) | Reconnaissance des erreurs dans un systeme de memoire | |
DE2157829C2 (de) | Anordnung zum Erkennen und Korrigieren von Fehlern in Binärdatenmustern | |
DE60215687T2 (de) | Fehlerkorrektion von multibit-baueinheiten mit unbeschränkter erkennung von doppelfehlern | |
DE102007007546A1 (de) | Fehlerkorrekturcode-Striping | |
DE1250163B (de) | Einrichtung zur Paritätsprüfung von Speicherworten | |
DE69508034T2 (de) | Fehlererkennungs-System und -Verfahren für gespiegelten Speicher zwischen doppelten Plattenspeichersteuerungen | |
EP0922251B1 (fr) | Systeme de memoire | |
DE2655653C2 (de) | Anordnung zur Feststellung der richtigen Zuordnung von Adresse und Speicherwort in einem wortorganisierten Datenspeicher | |
DE2538802C2 (de) | Schaltung zum Nachweis von Fehlern unter den aus Informations- und Prüfbits erzeugten, einen fehlerhaften Speicherort angebenden Bits | |
DE2325137A1 (de) | Speichereinrichtung mit bereitschaftsspeicherelementen | |
DE102018112584A1 (de) | Konfigurierbare Sensorvorrichtung und Verfahren zur Überwachung ihrer Konfiguration | |
DE102020111321A1 (de) | System und verfahren zur erkennung und berichtigung von datenfehlern eines transparenten registers | |
DE69924012T2 (de) | Verfahren und vorrichtung für speicherdata-fehlerdetektion und speichermodul-fehlerdetektion | |
EP0922252B1 (fr) | Reconnaissance des erreurs dans un systeme de memoire | |
EP0182134B1 (fr) | Méthode de mise en oeuvre d'un système multicalculateurs à sécurité intrinsèque comportant plusieur circuits inprimés d'entrées/sorties à sécurité non intrinsèque | |
DE69534316T2 (de) | Telekommunikationsanlage mit einem prozessorsystem und ein prozessorsystem | |
DE69813392T2 (de) | Verfahren zur Modul-Fehlerdetektion |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 19990219 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): DE FR SE |
|
GRAG | Despatch of communication of intention to grant |
Free format text: ORIGINAL CODE: EPIDOS AGRA |
|
17Q | First examination report despatched |
Effective date: 19990716 |
|
GRAG | Despatch of communication of intention to grant |
Free format text: ORIGINAL CODE: EPIDOS AGRA |
|
GRAH | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOS IGRA |
|
GRAH | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOS IGRA |
|
GRAA | (expected) grant |
Free format text: ORIGINAL CODE: 0009210 |
|
AK | Designated contracting states |
Kind code of ref document: B1 Designated state(s): DE FR SE |
|
REF | Corresponds to: |
Ref document number: 59701775 Country of ref document: DE Date of ref document: 20000629 |
|
ET | Fr: translation filed | ||
PLBE | No opposition filed within time limit |
Free format text: ORIGINAL CODE: 0009261 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT |
|
26N | No opposition filed | ||
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: SE Payment date: 20050811 Year of fee payment: 9 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: FR Payment date: 20050819 Year of fee payment: 9 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: SE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20060807 |
|
EUG | Se: european patent has lapsed | ||
REG | Reference to a national code |
Ref country code: FR Ref legal event code: ST Effective date: 20070430 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: DE Payment date: 20070831 Year of fee payment: 11 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: FR Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20060831 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20090303 |